๐ค AI Summary
In scientific machine learning, local explanations, while often accurate, frequently lack global consistency. To address this, this work proposes the SEAM frameworkโthe first approach to enable computable global explanation consistency across regions, sensors, and model components. SEAM models explanations via an explanation-sheaf formalism that integrates system states, closures, observation channels, and contract metadata. By leveraging channel-wise parsing obstructions and residual-aware regularization, it structurally assembles local explanations into globally coherent scientific interpretations, explicitly distinguishing between inconsistency and non-identifiability. Evaluated across 19 experiments, SEAM successfully identifies scenarios where local predictions are accurate yet explanations conflict, precisely attributing discrepancies to specific channels and overlapping regions, thereby supporting generator monitoring under distributional shift.
๐ Abstract
Scientific machine learning commonly validates models at the level of a subdomain, a benchmark split, or an explanation for one prediction. Yet such local checks cannot establish whether the resulting explanations can be assembled into one globally admissible explanation. We introduce Scientific Explanation-Admissibility Machines (SEAM), a generator-agnostic framework that makes this local-to-global consistency question computable across regions, sensors, regimes, and model components. The finite explanation-sheaf instantiation SEAM-$ฮฉ$ represents each region by a structured explanation with state, closure, and observation channels together with optional contract metadata; compares neighboring explanations on their overlaps; and converts disagreement into a channel-resolved obstruction. This obstruction locates inconsistency and tests competing declared accounts by restricting each repair to the revisions that one account permits. Exact feasibility refutes or retains an account; when exact repair is unavailable, residual-aware regularized records provide a separately labeled empirical attribution. The framework also separates inconsistency from non-identifiability and monitors learned generators under distribution shift. We establish theorems for minimum-cost intervention and conservation-contract detectability, together with companion results for identifiability and closure recoverability. Across nineteen experiments involving synthetic partial differential equation systems and out-of-distribution Fourier neural operator (FNO) monitoring, SEAM detects incompatible explanations even when local predictions are accurate, and attributes failures to specific channels and overlaps. SEAM adds a global explanation-consistency audit to existing solvers and learning models, testing whether their local explanations form a coherent scientific account.